COVID-19 “testing” is a word with a variety of meanings. We have to get it right.
We seem to encounter constant confusion or overly simplified descriptions of our evaluation strategies for COVID-19 as just “testing”… as if all and any tests are the same. But they are not. Similarly, we have not, on a per capita basis, tested more people than China, or South Korea or Italy. Hopefully this will change soon.
Some of our tests are gold standard and have been in short supply, high demand, and requiring days and weeks to obtain results. Others are rapid (hours or a few days) and even more recently, point-of-care with results known in minutes.
But all of these tests so far are genome tests sensing for viral DNA or RNA using a version of rt-PCR (reverse transcriptase, polymerase chain reaction). It identifies the presence of an antigen—in this case a specific infectious virus. PCR is a type of lab test that typically requires a high-level laboratory, highly trained technicians, known and validated probes and primers, good controls and elapsed time and precision to perform correctly.
At first, only CDC was allowed to conduct these tests with collected samples sent only to one of their sites, then other labs were certified, and now high-level university labs in most states with help from highly trained graduate students as extra technicians are racing to perform as many as possible because they have long been considered gold standard methods. The basic premise is that a sample of collected material via a nasal and/or pharyngeal swab will provide viral particles of a sort and amount which can be amplified and therefore detected if you have current COVID-19 infection.
However, swabs may not be collected correctly, nor enough material on each, and newer rapid diagnostic tests for COVID-19, like “quick flu” and “quick strep” have real and persistent issues of low sensitivity (false negatives) while typically being designed for higher specificity (true positives).
Estimates from China’s testing during the initial Wuhan crisis indicate swabs for their various rt-PCR tests were only 30-50% accurate. Even more alarmingly, only last week China released data indicating that even initially positive and hospitalized patients who seemed to recover symptomatically and had negative re-tests have tested later as positive. Positive, negative, negative, negative, positive! Theories abound on how this might be possible if there is no re-infection. This is shocking to consider. The last positive result suggests that many former patients deemed recovered, and presumably allowed to travel and congregate in large groups, are still shedding virus and possibly infecting others.
Was this because of an inaccurate set of prior tests or because new tests only detected non-infectious remaining viral particles? This is still not known. But another recently defined issue is that the China case count definitions are not what we might think either. We were told last week that there were no new cases in a certain region of China, but now we know that they do not include in overall daily counts for total and for new cases, any patients re-testing as positive after being released from the hospital or other medical care. These counts also do not include asymptomatic cases (but known positive) in the case counts we have been hearing. That is a very big issue, of course. Case definitions and test result definitions must be standardized if the world is to value and make decisions based upon them. “I have no idea why the authorities chose not to count [asymptomatic] cases in the official case count. I am baffled”, said one of the Wuhan doctors who had a second positive test after recovering. (NPR)
False negative rates will also be an issue with the latest rapid COVID test Abbott is running on the ID NOW platform and approved under EUA (emergency use authorization) by the FDA. But if we all eventually have access to this test under loosened inclusion protocols, I will try to get it.
It will be an important addition for quickly finding more active current infection, aka cases, albeit that experts agree false negative rates may run as high as 15% for COVID-19. This means that if you receive a negative result, you must continue to act and react as if you are still able to transmit the disease. THIS IS CRUCIAL TO UNDERSTAND! The faster test with results provided on-site will allow us to quickly know more about disease and transmission patterns, but still not all we need. It is likely that this consideration became clear to President Trump before he made his latest request for social distancing to continue through all of April. Observing just the strange pattern of negative to positive results in China would be enough…
But there may also have been a bit of good news in a recent Presidential briefing. With his administration finally agreeing to purchase test kits used so successfully by South Korea’s identification and containment efforts, we may soon also be able to include another very simple immune status test here in the US. “Soon” might still be a month away, but test kits have been purchased and submitted for FDA certification by at least one company. These types of tests are qualitative serology (blood) tests measuring antibody (not antigen) responses…IgM indicates current active infection, in this case with COVID-19, to which a person has mounted an antibody response. If such a person later is re-tested and shows IgG antibodies, it means they convalesced from the infection and are clear and immune for some unknown period; this also means they will no longer transmit the current viral strain. An initial IgG result implies similarly that you were at some point (likely relatively asymptomatically) infected and now recovered. Again, you would not infect or transmit the disease.
Of course there are finer points to antibody type, affinity, and sufficiency, but for what is known and needed at this time, showing an IgM vs convalescent IgG signal, or none of either (not as yet infected) would be of enormous interest for not only each person’s peace of mind, but also for national modeling efforts that could drive other decision-making for resources, care and interventions, as well as social and economic goals to follow..
So, antigen vs antibody testing? Longer PCR lead time vs rapid PCR? Nasal/pharyngeal swab vs finger stick/venous blood? My choice will be to get one of these when inclusion protocols begin to allow testing for less symptomatic people when more tests become available. Meanwhile we should stay sequestered for a while longer. You do not want to become seriously ill in the first wave of reduced resource availability.
I think we have a fortunate trajectory in our region, with fewer and later cases, lower modeled deficiencies forecast for beds and ICU interventions, and have been made safer earlier by swift mitigation efforts of our State leadership. We have a way to go; we are not China, but neither are we NY. So far, we are simply lucky.
Pamela Getson is a PhD biostatistician with a BS in physiology. She was Director of the Statistical Core Group of Children’s Hospital while on Faculty at the George Washington University School of Medicine/Pediatrics. She now lives in St. Michaels. In addition to teaching and research, she designed and evaluated clinical trials, and later served as a senior math/biostat reviewer of data submitted in support of new vaccines at the FDA/CBER. She consults to industry and academia, but now living on the shore permanently in retirement, primarily assists with her husband’s new interest in oyster gardening and resumption of their love of sailing and nature.
Lois Shepard says
Thank you for this Excellent, Informed, and Important article.
Willard Engelskirchen says
A very good explanation of a complicated topic. A decision flow chart would be helpful here but also complicated. The bottom line is this: Stay home, read what you can, try to understand this, and TAKE IT SERIOUSLY.
Thank you.